[PATCH] D54740: [NewPM] fixing asserts on deleted loop in -print-after-all

[Philip Pfaffe via Phabricator via llvm-commits]

philip.pfaffe added a comment.

> Not really. If you say have a Loop* pointer, which is freed and then another Loop allocated happens to be just at that pointer
>  then you dont have a way to identify whether it is the old Loop or the new one.
>  Presumably that does not happen now...

The point is what kind of information you attach to the handle. When triggering the invalidation callbacks, the message is "This particular loop was invalidated". And that's a valid message, no matter if the handle is already being reused, because a downstream instrumentation cannot have observed the new loop behind the handle yet since we're still on the invalidation path.


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D54740/new/

https://reviews.llvm.org/D54740